Summary The Jetsgo DC-9-83 C-FRYH, serial number 53520, was operating as JGO191 on a flight from Toronto/ LesterB. Pearson International Airport, Ontario, to Calgary International Airport, Alberta. The runway visual range passed to the crew with the landing clearance for Runway34 was 1400feet, with a runway light setting of five. JGO191 conducted the instrument landing system (ILS) approach and touched down on the runway, left of the centreline at 1956 mountain standard time (MST). The aircraft departed the left side of the runway surface and travelled 1600feet before climbing out on a missed approach procedure. A hold short sign was struck and destroyed while the aircraft was on the ground. JGO191was vectored back to Runway34 for a second ILSapproach and landed at 2010MST. There was minor damage to the aircraft, and there were no injuries among the 78passengers and 6crew members. Ce rapport est galement disponible en franais. Other Factual Information The flight crew were licensed in accordance with the existing Canadian Aviation Regulations (CARs). The pilots' flight duty days were within the prescribed regulations and both were rested. A special weather observation taken at the Calgary International Airport at 1932 mountain standard time1 was as follows: wind 040True(T) at 6knots; visibility statute mile (sm) in light snow and freezing fog; runway visual range (RVR) for Runway34, 2800feet; ceiling 100feet above ground level (agl). In the remarks section, fog was reported to have 8oktas2 sky coverage. The 2000weather observation was: wind 040T at 7knots; visibility sm in freezing fog; ceiling overcast at 400feet agl; temperature -4C; dew point -6C; altimeter 29.74. In the remarks section, fog was at 6oktas and stratus clouds were at 2oktas. The RVR for Runway34 was 1400feet. Radar vectors were given to JGO191 to position the aircraft for the CatI instrument landing system (ILS) 34approach. Approach clearance was given by Calgary Arrival with an RVR value of 1600feet. Initial contact with Calgary Tower indicated that the RVR for Runway34 was 1600feet. Calgary Tower cleared JGO191to land with the advisory that the preceding aircraft had reported sighting the approach lights at decision height (DH) (200feet agl), and that the current RVR value was 1400feet. The approach was flown in accordance with the existing CARs. The captain, as the pilot flying (PF), flew from the left seat. The autopilot was engaged and coupled to the ILS receiver, and the auto throttles were engaged, which provided a stabilized approach profile at about 135knots. The pilot not flying (PNF) monitored the approach from his instruments and called 100feet above DHand thenDH. Periodically, the PNF looked outside the cockpit to identify the runway lights. The PFflew the aircraft with reference to the flight instruments until reaching theDH, where, upon gaining visual reference to the approach lights, he elected to land the aircraft. At DH (1954:55), the captain looked up and called for a landing. The autopilot was disconnected at 1955:05, approximately 65feet above aerodrome elevation. The PF rolled into a 10bank to the left, resulting in a heading change of 5to the left. The PNFnoticed that the aircraft was drifting to the left and advised thePF. The PFacknowledged the PNFand continued with the landing. The aircraft touched down on the runway at 1955:13, approximately 1400feet from the threshold of Runway34 and about 80feet left of the centreline, on a track of about 335Magnetic(M) (see Figure1). The aircraft departed the left side of the runway 1600feet from the threshold, eventually parallelling the runway about 71feet from the edge. Take-off power was applied to initiate a go-around at 1955:17. While travelling through the grass, the aircraft's left outer flap struck a glancing blow to the hold short sign located south of Runway07/25. JGO191 proceeded across Runway07/25 and struck and destroyed the hold short sign north of Runway07/25 with the left main landing gear before lifting off at 1955:25. The aircraft was on the ground for approximately 1800feet and 11.5seconds. Overhead view of the runway - Calgary International Airport After the climb out, the crew observed the "Red Light On With Landing Gear Handle Up" indication, which indicated that the landing gear had not successfully retracted. The appropriate checklist items were reviewed and it was determined that, due to undetermined damage to the landing gear, the aircraft would be restricted to 230knots. The crew elected to conduct a second approach into Calgary due to the speed restriction and the high prevailing winds from the west that would preclude the use of Abbottsford as an alternate. JGO191 was advised by air traffic services that there were nine aircraft holding ahead and if they could not accept an extended downwind they should plan for their alternate or declare a fuel emergency. JGO191 declared a fuel emergency, with 11000pounds of fuel on board, in order to be sequenced to the front of the line. A second ILS approach to Runway34 was successfully executed. The RVR value on Runway34 at the time of the second approach was 1400feet. Runway 34 was equipped with high intensity runway edge lighting, with variable settings to a maximum of five. The lighting system was comprised of white parallel runway edge lights, green threshold lights, red runway end lights, 1000feet of sequenced flashing approach lights, and 1400feet of white runway alignment indicator lights. Runway 34 was served by a precision CatI ILS approach. The DH was 200feet agl with an advisory visibility of mile or 2600feet. The ILS provided a standard 3glide slope on a final approach course of 343M. A NAV CANADA Navigational Status Report taken at the time of the occurrence indicated that the ILS readings for Runway34 were normal. The last flight check of the ILS was conducted on 15May2004; the results were within normal parameters. Runway 34 had an RVR typeA sensor located adjacent to the runway threshold. The sensor measures light emitted from its projector to calculate the maximum horizon distance that may be seen from a point above its centreline at a height corresponding to the average eye level of pilots at touchdown.3 The amount of light detected may be attenuated by snow, fog and rain. The sensor was subject to monthly maintenance inspections. Prior to the occurrence, the sensor was inspected on 27December2004 and found to be accurate. Runways serviced with an RVR are subject to an approach ban (CARs602.129). The regulation provides the only visibility-based restriction for the conduct of an approach or landing. This allows the pilot to conduct an approach to a runway anytime the RVR is at least 1200feet, or in cases where the RVR is not available or not provided. Visibility values provided in the Canada Air Pilot (CAP) approach plates are advisory only and, if prevailing at the time of approach, should result in the required visual reference being established and maintained to landing. A pilot may continue with a landing provided that visual contact with the runway environment was made prior to passing the minimum descent altitude or DH (see CAPGEN13). Riordan (1974) describes the process through which the pilot acquires and interprets the visual cues available in the runway environment.4 To assess glide slope and runway closure, the pilot will: assess runway perspective: the pilot assesses the size of the runway (which varies with horizontal distance from the runway) and shape of the runway (which varies with vertical distance from the runway); assess visual rate of change: the pilot assesses the rate of change of these two variables to provide information with respect to the horizontal closure rate (change of apparent runway size) and vertical position with respect to the glide slope (change of apparent runway shape); and, assess runway motion parallax: the pilot uses any apparent motion of the target touchdown point to provide an indication of rate of descent. Available visual segment at 200feet and at 100feet on an approach When pilots are making the transition from instrument to visual flight, their ability to judge the approach, flare and landing rollout will be dependent upon the length of the visual segment (that is, the distance ahead of the aircraft that the pilot can see). At a DHof 200feet, on a 3glide slope, the aircraft is 3816feet from the touchdown zone (3816'=200Tan 3). Carmack(1972)5 points out that the downward vision angle of most transport aircraft is approximately 14, which will significantly reduce the length of the visual segment. At a height of 200feetagl, approximately 800feet (800=200Tan14) in front of the aircraft will be obscured by the nose of the aircraft, and at 100feetagl, the amount of obscured terrain will be approximately 400feet (400=100Tan14). Given a 1400-foot RVRand assuming this to be representative of slant range visibility, this will provide a visual segment at a 200-foot DHof approximately 600feet, and a visual segment of approximately 1000feet at 100feetagl. Carmack summarized the effect of this visual segment as follows: "A visual segment of 800feet is insufficient to consider as visual conditions, so the aircraft must be flown by instruments below decision height."6 In addition to assessing the cues described above, to judge rate of closure and rate of descent, the pilot must have sufficient cues to assess any lateral deviation from the runway centreline and correct for drift. The time required to complete the transition from instruments to visual cues was assessed during the Carmack study. The following is a conclusion from that study: It was determined it requires about three seconds for the heads-down pilot to integrate the outside visual cues after becoming visual. It requires this length of time to adjust to the outside environment, to determine position with relation to the runway, determine cross-track rate and develop the knowledge to effect the control inputs necessary for visual control.7 The time required for this transition will be increased in situations where the aircraft is offset from the runway centreline at minimums (e.g. during crosswind approaches), or where obscured or partially obscured conditions as visual cues will become indistinct and easily lost in these conditions.8 Any hesitation in making the transition from inside to outside references will also vastly increase the time required to fully integrate visual cues. Studies have demonstrated that a pilot requires a minimum of 700milliseconds to transition from outside references to inside references. During this period, the crew member is not capable of attending to information either inside or outside of the cockpit.9 Therefore, a pilot faced with indistinct visual cues attempting to compensate with reference to the flight instruments will have their cognitive resources taxed significantly during the last moments of the approach. Hoglund (1982)10 indicates that the time available for decision making between a 200-foot DHand the threshold on a 3glide slope would be 19seconds for an aircraft travelling at 120knots and 16seconds for an aircraft travelling at 140knots. Based on this and the lighting systems available, he recommended "look-see" approaches be prohibited if RVR values were less than 1800feet for CatI operations. Young (2003)11 indicates that a particularly [critical] period for the pilot occurs when making the transition from instrument flight to flying by external visual cues. There is not a specific illusion associated with the transition, but rather a period of uncertainty concerning orientation. A pilot who has been concentrating on the instruments in lining up for a landing may easily experience spatial disorientation during the several seconds after looking up and trying to find the runway and the horizon through broken clouds. A vection (visually induced perception of self-motion) illusion is produced by the nearly uniform motion of a large part of the visual field, resulting in the person feeling that they are moving opposite to the motion observed in the visual field.12 Therefore, fog moving across a runway could give a pilot a false sense of drift. However, this illusion will be relatively easily overcome if sufficient stationary cues are available, such as additional lighting. In this occurrence, as described above, the available visual cues were marginal to allow the landing to be carried out effectively, even in the absence of a vection illusion. The presence of moving elements in the visual scene, such as blowing fog or snow, would only compound the problem. A number of occurrences have been investigated by the TSB in which inadequate visual references during the final stages of an approach contributed to an accident. A91A0198: A DC-8 aircraft was conducting an ILS approach to Runway29 at the Moncton Airport, with a reported RVR of 1400feet. After touchdown, the aircraft left the right side of the runway and travelled approximately 1100feet prior to regaining the runway. During the final stages of the landing, the crew had difficulty discerning the runway. A93W0037: A Boeing 737 conducting a CategoryI ILSto Runway16 at the Calgary Airport departed the left side of the runway after touchdown. The aircraft struck a number of runway and taxiway lights prior to regaining the runway surface. The weather at the time of the occurrence was a ceiling of 200feet obscured and a visibility of 1/8mile in very light freezing drizzle and fog. The RVR was reported to the crew as 2400feet; runway lights were switched to a setting of five just prior to commencing the approach. A97H0011: A Canadair CL 600 Regional Jet was conducting a CategoryI ILSapproach to Runway15 at the Fredericton Airport. The crew had sight of the approach lights at minimums and elected to land. On arrival, the reported ceiling was 100feet obscured, the visibility 1/8mile in fog, and the RVR 1200feet. On reaching about 35feet, the captain assessed that the aircraft was not in a position to land safely, as it was left of the centreline and the crew had no way to assess how far down the runway they were; the captain ordered a go-around. As the aircraft reached its go-around pitch attitude of about 10, the aircraft stalled aerodynamically and impacted the ground. A99Q0151: A Raytheon Beech 1900D was on a scheduled flight from Port-Menier to Sept-les, Quebec, with two pilots and two passengers on board. The aircraft crashed while on approach to the airport, 1nautical mile (nm) short of the runway, in reported weather conditions of a 200-foot ceiling and sm visibility. The crew had descended well below safe minimum altitude while in instrument meteorological conditions. A03Q0151: A PA-31-310, with one pilot and two passengers on board, was on a visual flight rules flight from les-de-la-Madeleine, Quebec, to Gasp, Quebec. While en route to Gasp, the pilot was informed that weather conditions at his destination were a ceiling of 500feet and visibility of sm in fog. The wreckage was found on a hilltop 1.2nm northeast of the airport. The pilot had continued his descent below minimum descent altitude without having the visual references required to continue the landing. A04W0032: A Boeing 737 aircraft was conducting an ILSapproach to Runway15 in Edmonton, in conditions of freezing fog with a reported RVR of 1200feet; runway lights were at a setting of five. The aircraft touched down to the left of the runway and travelled approximately 1600feet before returning to the runway. These occurrences share a number of commonalities. All were conducted during darkness with visibilities less than those recommended on the CAPapproach plate to runways served by a CategoryI ILS system. In these occurrences, the crew had sight of the runway environment at minimums and elected to land, but subsequently had difficulty acquiring sufficient visual references to maintain aircraft alignment with the runway. The Transportation Safety Board has identified safety deficiencies associated with conducting approaches in low visibilities. As a result of investigation A97H0011, TSB Recommendation A99-05 was issued, advising the following: The Department of Transport reassess CategoryI approach and landing criteria (re-aligning weather minima with operating requirements) to ensure a level of safety consistent with CategoryII criteria. In its 06 August 1999 response to TSB Recommendation A99-05, Transport Canada indicated that draft regulatory amendments to strengthen the standards for instrument low-weather approaches would be submitted without delay to the Canadian Aviation Regulatory Advisory Council for consultation, with the goal of implementing changes as soon as possible. On 12 August 1999, another accident occurred where approach visibilities were identified as an underlying factor (A99Q0151). As a result of the investigation, TSB Recommendation A02-01 was issued, which stated in part: From January 1994 to December2001, the Board has investigated 24such accidents where low visibilities and/or ceilings likely contributed to the accident. These accidents resulted in 34fatalities and 28serious injuries, not counting the loss of property and damage to the environment. In September1999, TC initiated action to implement new approach ban regulations based on visibility.... This process has been ongoing for two years now; however, its timely implementation has been delayed because of some resistance. The Department of Transport expedite the approach ban regulations prohibiting pilots from conducting approaches in visibility conditions that are not adequate for the approach to be conducted safely. In its 26 May 2002 response, Transport Canada indicated that it had prepared 16Notices of Proposed Amendments (NPAs) to address TSB Recommendation A02-01, concerning approach ban regulations related to visibility. These NPAs were under review at the Department of Justice at the time, and it was expected that the final product would be published in the June2002 issue of the Canada Gazette. On 25 November 2004, the Board released report A03Q0151, which identified that Transport Canada's proposed approach ban regulatory initiative should decrease the probability of accidents on instrument approaches in reduced visibility conditions. However, the report further stated that, "The Board was nonetheless concerned that, until these proposed regulatory provisions come into force, safety measures will remain inadequate against the risk of controlled flight into terrain resulting in loss of life." The drafted NPAs were published in the Canada Gazette, PartI, in November2004 for comment. The period for comment ended in January2005. Transport Canada is in the process of reviewing these comments and the Board will continue to monitor the progress of these amendments. The proposed changes to the approach ban regulations would include an increase from RVR 1200to RVR 1800for CategoryI precision approaches servicing runways without centreline lighting for commercial operations. For airline operations (CAR705), authorized through an Operations Specification requiring additional training requirements and aircraft and aerodrome equipment requirements, the required RVR would be 1200for a HUD (head-up display) equipped aircraft, or 1600for a non-HUD equipped aircraft.